Machine learning - Logistic regression

Рет қаралды 39,139

Күн бұрын

Logistic regression: Optimization and Bayesian inference via Monte Carlo.
Slides available at: www.cs.ubc.ca/~nando/540-2013/...
Course taught in 2013 at UBC by Nando de Freitas

Пікірлер: 14

@harry1357931gmail 11 жыл бұрын

Excellent Lecture on Logistic Regression....must go through it twice to thrice to absorb it completely....

@mrf145 9 жыл бұрын

Thank you. The lecture is quite easy to understand. You delivered it in very good way.

@18amarage 8 жыл бұрын

I really love your lecture sir...

@brianclark4796 10 жыл бұрын

thank you for making this lecture available. very helpful.

@anynamecanbeuse 4 жыл бұрын

This is just blowing my mind I should say.

@vermajiutube 6 жыл бұрын

Awesome lecture. How do we estimate w(thetha_i) using MC in bayesian ?

@karthiks3239 10 жыл бұрын

This was a nice lecture.. Thank you.. I was also looking for lectures on Constrained Optimization and Support Vector Machines.. Are these available?

@flamingxombie 7 жыл бұрын

Great lecture. By 'simulating' thetas, I am reckoning that we are respawning thetas from regions in proportion to their probabilities?

@abdulrahmansattar2873 6 жыл бұрын

Superlike!

@JackSPk 5 жыл бұрын

Maybe I'm misunderstading the 3D plot, but shouldn't it be Y=0={RED} and Y=1={BLUE} at 12:40 ?

@Jinex2010 8 жыл бұрын

22:00 I think J comes Jacobian and H from Hessian

@Raven-bi3xn 3 жыл бұрын

But the question is about the J of cost function.

@Romis008 6 жыл бұрын

The prof repeatedly mentions that entropy is the opposite of information... I always thought entropy was expected shannon information content. Am I missing something? By the way, amazing content!

@Cindy-md1dm 5 жыл бұрын

Entropy measures the uncertainty, it is negative of information. When the probability goes close to 1, which means more information you get, less uncertainty left.